Product Changelog - ISMS Copilot updates

30 March 2026

Improvements

Excel Layout Preservation

XLSX files now receive special handling to preserve spreadsheet structure during upload and analysis. Tables keep their columns, rows, and orientation intact instead of being flattened or transposed. Risk registers, control matrices, and asset inventories uploaded as Excel files now appear with their layout closer to the original.

What's improved:

• Table structure preserved — Columns, rows, and cell relationships stay as they were in your original file

• Orientation maintained — Vertical data stays vertical, horizontal stays horizontal

• Wide tables supported — Spreadsheets with many columns convert properly through a dedicated table-aware path

This improvement is automatic — upload XLSX files as usual and the system routes them through the table-preserving conversion path. No configuration needed.

For best results with complex spreadsheets, ensure your file is under 5MB and uses clear table headers. The AI reads both the structure and content to provide accurate compliance analysis.

Overview

Stay up to date with the latest features, improvements, and bug fixes for ISMS Copilot. This changelog is automatically updated as we ship new changes to production.

26 March 2026

New Features

Team Workspaces

Invite team members to collaborate on shared workspaces. Create a team from Settings → Team, invite by email, and share workspaces with your team. Each member gets their own usage limits at the same plan price (per-seat billing). Shared workspaces show a “Shared” badge in the workspace selector.

Data Export

Download all your personal data in a machine-readable JSON file from Settings → Data Protection. Includes conversations, files metadata, workspaces, memories, and settings. Compliant with GDPR Article 20 (right to data portability).

Account Self-Deletion

You can now permanently delete your account and all associated data from Settings → Data Protection. If you have an active subscription, you’ll be asked to cancel it first. This action is irreversible and removes all conversations, files, documents, workspaces, memories, and settings.

Safety

Content Moderation

All chat messages are now automatically screened for harmful content (hate speech, violence, self-harm, etc.) using Mistral’s moderation model. Flagged conversations are protected from deletion to support audit and review.

25 March 2026

New Features

Inline Memory Editing

Click the pencil icon on any memory to edit it in place. No more delete-and-re-add when facts change.

Memory Source Badge

Each memory now shows a small “auto” or “manual” label so you can tell which were detected by the AI vs typed by you.

Improvements

Memory Limit Doubled

Memory capacity raised from 50 to 100 memories per workspace and per user. Context injection cap also doubled (4K → 8K chars), allowing the AI to reference more of your stored context in each conversation.

21 March 2026

New Features

Improved Advanced Data Protection

The EU-based AI model now supports reasoning mode (think) and has a larger context window (256K tokens). Provider names have been removed from user-facing text for cleaner, provider-agnostic messaging.

Memory Capacity Indicator

See how many memories you have (e.g., “3/100”) in both Settings and workspace views. The Add button is disabled when full, and a one-time notification explains that new facts won’t be saved until you free up space.

Bug Fixes

Fixed Generated Documents Cut Off at ~4K Characters

Downloaded DOCX/PDF files were missing most of their content (e.g., only sections 1–3 out of 11). The AI formatting step now uses a model that reliably produces complete output.

18 March 2026

New Features

Memories for General Conversations

Memories now also work outside of workspaces. Personal facts you share in general conversations — your role, preferences, compliance scope — are remembered across sessions.

Memory Toggle

Enable or disable automatic memory detection per scope in Settings → Memories. Toggle general and workspace memories independently. Disabling keeps existing memories but stops new ones from being created or used.

Bug Fixes

Better Error Handling When AI Service Is Busy

If the AI service is temporarily overloaded during a response, you now keep the partial response you already received instead of losing it. The error message is clearer (“AI service is temporarily busy”) and uses a softer amber warning instead of a red error.

17 March 2026

New Features

Advanced Data Protection Badge

When Advanced Data Protection is enabled, a blue shield badge now appears in the chat input area. Hover over it to see that you're using Mistral (EU-hosted, GDPR-compliant) and get a reminder that you can switch back for more demanding tasks.

Improvements

Improved Document Generation

PDF and DOCX exports now use a markdown-based pipeline instead of AI-generated HTML. This produces cleaner formatting and eliminates truncation, even for 20+ page policies covering all 93 Annex A controls.

16 March 2026

New Features

Workspace Memories

The AI now remembers your organization's context — tools, team size, certifications — across all conversations in a workspace. Memories are auto-detected from your messages and can also be added manually. Each workspace maintains its own separate memory, so context from one client or project never leaks into another.

Bug Fixes

Fixed False Subscription Downgrades

Fixed an issue where Stripe webhook race conditions could cause paying users to be incorrectly downgraded. Subscription state is now handled atomically.

Fixed Document Truncation

Long generated documents are no longer silently cut off during export.

14 March 2026

New Features

Framework Detection Display

When ISMS Copilot detects relevant compliance frameworks during a conversation, they are now visibly displayed in the chat interface. You can also see which version of each framework the AI is using.

12 March 2026

Improvements

Smarter Document Exports

Document generation now uses Claude Sonnet 4.6 for higher-quality outputs. Max output tokens raised from 8K to 64K, allowing much longer and more detailed responses — fixing issues where long exports (e.g. detailed role descriptions) were cut off.

File Upload Quota Fix

Only fully processed files now count toward your monthly upload limit. Failed or stuck uploads no longer reduce your quota.

Refreshed Login Page

The login and signup page has a fresh new wallpaper.

11 March 2026

New Features

Custom Instructions

You can now set custom instructions that shape how ISMS Copilot responds to you. Define your preferred language, tone, compliance focus, or any other guidance — the AI will follow it across all your conversations.

UK Data Protection Frameworks

ISMS Copilot now supports three UK-specific frameworks:

• UK GDPR — retained EU GDPR under UK law (ICO as supervisory authority)

• UK Data Protection Act 2018 — all Parts covering general, law enforcement, and intelligence services processing

• UK Data Use and Access Bill — marked as draft/emerging legislation

10 March 2026

New Features

Document Library

All your generated documents are now accessible from a single dedicated page. Browse, filter, and manage every PDF, DOCX, XLSX, and Markdown file you've created across all workspaces and general conversations.

What's New:

• Unified View: See all generated documents in one place, accessible from the sidebar

• Filter by Format: Toggle between PDF, DOCX, XLSX, and MD format chips to find what you need

• Filter by Workspace: Tab bar lets you view documents from All, General, or specific workspaces

• Sort by Date: Toggle between newest and oldest first

• Full Document Actions: Preview, download, rename, or delete any document directly from the library

Per-Plan File Upload Limits

File upload limits are now tailored to your subscription plan, giving paid users significantly more capacity.

Monthly Upload Limits:

• Free: 5 uploads/month

• Plus: 50 uploads/month

• Standard: 150 uploads/month

• Pro: 300 uploads/month

• Business: 500 uploads/month

Failed file uploads no longer count against your monthly quota. The upgrade dialog now shows the specific upload limit for each plan so you can compare easily.

Reset Stuck Conversations

If a conversation gets stuck while the AI is processing, you can now reset it yourself instead of waiting or losing the thread.

How It Works:

• Automatic Detection: If a conversation has been stuck for more than 2 minutes, an amber banner appears with a "Reset thread" button

• One-Click Fix: Click the button to unlock the conversation and resume chatting

• Automatic Recovery: Conversations stuck for more than 15 minutes are now automatically recovered in the background

Improvements

AI Responds in Your Language

The AI assistant now respects your interface language setting when responding. If you've set your interface to French, German, Spanish, Italian, or Dutch, the AI will respond in that language automatically.

Service Questions Redirected to Help Center

When you ask the AI about ISMS Copilot pricing, billing, licensing, or account features, it now directs you to the Help Center at help.ismscopilot.com for accurate, up-to-date information instead of answering from potentially outdated knowledge.

Feedback Forms Now Include Context

Feedback and issue report forms now include conversation context for better issue triage.

Bug Fixes

Subscription Loading Errors No Longer Silent

Fixed an issue where a temporary database error could silently downgrade paid users to free-tier limits. Errors now show a clear notification so you can reload and retry.

Improved File Processing Security

Strengthened authentication and authorization checks on the file processing endpoint to ensure files can only be processed by their owner.

File Content No Longer Reloaded After Compaction

File content is no longer unnecessarily reloaded after conversation compaction, improving performance in long conversations with attachments.

7 March 2026

New Features

New Standard Plan

A new Standard subscription tier is now available, sitting between Plus and Pro. It gives you more credits per session than Plus at a lower price point than Pro—ideal if you need more than the basics but don't require the full Pro allowance.

What's New:

• Standard Tier: A mid-range plan with more generous session credits than Plus

• Updated Upgrade Page: The plan comparison grid now shows all four tiers side-by-side so you can easily compare what each plan offers

Sort Workspaces

You can now sort your workspace list to find what you need faster.

Sorting Options:

• Alphabetical (A→Z): Sort workspaces by name

• Newest First: See recently created workspaces at the top

• Oldest First: See your longest-running workspaces first

Search Conversations (Ctrl+K)

Find past conversations instantly with full-text search across all your chat history. The new search dialog lets you search through message content and conversation titles to quickly locate specific compliance discussions, control implementations, or policy reviews.

What's New:

• Keyboard Shortcut: Press Ctrl+K (Cmd+K on Mac) anywhere to open the search dialog

• Full-Text Search: Search across both conversation titles and message content for comprehensive results

• Highlighted Snippets: See matching text highlighted in context to quickly find the right conversation

• Workspace Filtering: Search is scoped to your current workspace by default

• Recent Conversations: When idle, the dialog shows your most recent chats grouped by date

How to Use:

1. Press Ctrl+K (or Cmd+K) or click "Search chats" in the sidebar

2. Start typing to search your conversation history

3. Click any result to open that conversation

4. Press Escape to close the dialog

Link Conversations to Workspaces

Organize your work better by linking any conversation to a workspace. The new attach menu lets you associate chats with client workspaces even after they've started, making it easier to keep compliance work organized.

What's New:

• Plus (+) Button: Click the + icon next to file upload to link a conversation to a workspace

• Link Existing Chats: Attach ongoing conversations to workspaces retroactively

• Workspace Badge: See which workspace a conversation belongs to with an inline chip

• Quick Unlink: Remove workspace associations by clicking the X on the workspace chip

• Available Everywhere: Works in both the main chat and when creating new conversations

How to Use:

1. Click the + icon next to the paperclip (file upload) in the chat input

2. Select a workspace from the dropdown list

3. The conversation is immediately linked — you'll see a workspace chip appear

4. To unlink, click the X on the workspace chip

Auto-Expanding Message Input

The chat input now grows automatically as you type longer prompts, making it easier to review and edit detailed compliance queries before sending.

What's Improved:

• Smart Growth: Input box expands as you type, up to 50% of screen height

• Internal Scrolling: Beyond 50vh, the input scrolls internally so messages stay visible

• Message Anchoring: Chat messages stay anchored to the bottom as the input grows

• Works Everywhere: Applies to all chat input areas across the platform

No more squinting at a tiny input box when crafting multi-paragraph prompts or pasting long policy excerpts for review.

Affiliate Program

Refer others to ISMS Copilot and earn rewards through our new GDPR-compliant affiliate program.

Improvements

Usage Now Displayed in Credits

Your session usage is now shown in credits instead of raw tokens, making it much easier to understand how much of your plan allowance you've used.

What Changed:

• Credits Instead of Tokens: Settings → Usage now shows credits used and credits remaining, mapped to your plan tier

• Clearer Limits: Per-session credit limits are displayed on the upgrade page so you can compare plans at a glance

• Plan-Agnostic Messaging: Rate limit messages now guide you to upgrade without referencing a specific plan name

Improved Handling of Large Conversations and File Uploads

Long chat threads and large file uploads are now handled more gracefully, preventing freezes and unclear errors.

What Changed:

• Pre-Flight Size Check: Before sending your message to the AI, the system now checks whether the combined conversation size exceeds the limit—and shows a clear error immediately instead of timing out

• Smarter File Processing: Large Excel files with many empty rows are now cleaned up automatically, and very large extracted file content is trimmed to prevent slowdowns

• Long Thread Protection: Oversized historical messages in long conversations are automatically trimmed so the conversation can continue without hitting limits

Improved Caching

Better cache breakpoint handling for faster AI responses.

Smarter Conversation Compaction

Compaction trigger lowered from 150K to 100K tokens for smoother long conversations.

Better Expired Link Handling

Password reset and magic login links now show clear, user-friendly error messages when they expire, instead of silently failing or showing generic errors.

What's Improved:

• Clear Error Messages: "Email link expired" message appears when you click an old password reset or magic link

• Covers Both Flows: Works for both password reset links and magic (passwordless) login links

• Localized Errors: Error messages appear in your selected interface language

• Consistent Experience: Same clear messaging whether the error comes from the URL hash or query params

If you see this error, simply request a new password reset or magic link from the login page — the old link has expired for security.

Bug Fixes

File Size Limits Enforced Correctly

File upload limits are now properly enforced to prevent processing failures. Files requiring conversion (PDF, DOCX, XLS, XLSX) are limited to 5MB to avoid stuck uploads due to Edge Function memory constraints.

What's Fixed:

• Convertible Files (PDF, DOC, DOCX, XLS, XLSX): 5MB limit prevents timeout errors

• Simple Files (TXT, CSV, JSON): Retain the 10MB limit

• Clear Error Messages: You'll see specific feedback if a file exceeds the limit

• Updated UI: Drag-and-drop overlay now shows correct limits for each file type

28 February 2026

New Features

Upgrade Celebration

New users who upgrade from free to paid plans now see a delightful confetti celebration acknowledging their upgrade.

What's New:

• Confetti Animation: Fun visual celebration appears after successful payment

• Benefit Highlights: Popup reminds you of the premium features you've unlocked

• Smart Triggers: Shows when returning to Stripe after payment or revisiting threads where you previously hit rate limits

• Localized: Celebration messages appear in all 6 supported languages

Copy Entire Conversation

Export full conversations as Markdown with a single click. The new conversation menu in the top bar lets you copy entire chat histories for documentation, sharing, or archiving.

What's New:

• Conversation Menu: Three-dot menu button in the top bar (next to ghost mode toggle)

• Copy as Markdown: One-click export of the full conversation including all user messages and AI responses

• Formatted Output: Clean Markdown format with proper heading hierarchy and message attribution

• Available in Temporary Chats: Works in both saved conversations and temporary chat sessions

How to Use: Click the three-dot menu (⋯) in the top bar and select "Copy Conversation." The entire chat history is copied to your clipboard as Markdown, ready to paste into documentation tools, wikis, or notes.

User Activity Statistics

Track your productivity with the new Stats section in Settings. See how much time you've saved and how actively you're using the platform.

What You'll See:

• Messages Sent: Total number of AI interactions

• Conversations: Number of distinct chat threads

• Documents Generated: Count of policies, procedures, and compliance documents created

• Hours Saved: Estimated time saved based on your activity (calculated at ~10 minutes per AI response)

How to Access: Open Settings and click the Stats tab to view your personal usage metrics.

Improved Table Copy Formatting

Tables in AI responses now copy as properly formatted HTML instead of raw Markdown, making them paste cleanly into Word, Google Docs, email clients, and other rich text editors.

What's New:

• Rich Text Clipboard: Copy button now writes both HTML (for rich text editors) and plain text (for code/markdown editors) to clipboard

• Clean Paste Experience: Tables paste as actual tables in Word, Google Docs, Outlook, Gmail

• Styled Tables in Chat: Tables in the chat interface now have proper borders, padding, and header styling

• Automatic Fallback: Browsers without rich clipboard support still get markdown as plain text

No more manually reformatting tables after copying from ISMS Copilot. Control mappings, gap analysis tables, and compliance checklists now paste ready-to-use.

In-App Password Change

You can now change your password directly in Settings without leaving the platform. The new password change interface includes real-time strength feedback to help you create secure credentials.

What's New:

• Settings Integration: Access password change from Settings → Account or Security

• Strength Indicator: Visual feedback shows password strength as you type (weak, medium, strong)

• Validation Guidance: Clear requirements for minimum length, character types, and complexity

• Secure Process: Requires current password confirmation before allowing changes

How to Change Your Password:

1. Navigate to Settings and find the password change section

2. Enter your current password

3. Create a new password (minimum 8 characters recommended)

4. Watch the strength indicator to ensure adequate security

5. Confirm your new password and save

Usage Tracking in Settings

Monitor your token usage in real-time with the new Usage section in Settings. See exactly how much of your 4-hour session quota you've used, when it resets, and which plan you're on.

What's New:

• Usage Tab: New dedicated section in Settings between General and Data Protection

• Progress Bar: Visual indicator shows token consumption as a percentage of your plan limit

• Token Count: Clear display of used vs. limit (e.g., "45,000 / 300,000 tokens")

• Reset Time: Shows when your quota resets (e.g., "Resets at 16:00 UTC")

• Plan Info: Displays your current plan and upgrade link if applicable

How to Access: Open Settings (gear icon) and click the Usage tab. Your current session usage appears with a clean progress indicator.

User Activity Statistics

Track your productivity with the new Stats section in Settings. See how much time you've saved and how actively you're using the platform.

What You'll See:

• Messages Sent: Total number of AI interactions

• Conversations: Number of distinct chat threads

• Documents Generated: Count of policies, procedures, and compliance documents created

• Hours Saved: Estimated time saved based on your activity (calculated at ~10 minutes per AI response)

How to Access: Open Settings and click the Stats tab to view your personal usage metrics.

Copy Entire Conversation

Export full conversations as Markdown with a single click. The new conversation menu in the top bar lets you copy entire chat histories for documentation, sharing, or archiving.

What's New:

• Conversation Menu: Three-dot menu button in the top bar (next to ghost mode toggle)

• Copy as Markdown: One-click export of the full conversation including all user messages and AI responses

• Formatted Output: Clean Markdown format with proper heading hierarchy and message attribution

• Available in Temporary Chats: Works in both saved conversations and temporary chat sessions

How to Use: Click the three-dot menu (⋯) in the top bar and select "Copy Conversation." The entire chat history is copied to your clipboard as Markdown, ready to paste into documentation tools, wikis, or notes.

Automatic Conversation Compaction for Think Mode

Long conversations in Think mode now continue indefinitely without hitting token limits. Server-side compaction automatically summarizes earlier portions of extended chats, allowing you to maintain context across unlimited exchanges.

What's New:

• Infinite Conversations: Think mode chats now continue seamlessly past the typical ~150,000 token threshold

• Automatic Summarization: Backend automatically compacts older messages when approaching context limits, preserving conversation flow

• Visual Indicator: Brief "Compacting our conversation..." progress indicator appears during background summarization

• Transparent Process: Compaction happens seamlessly—your conversation resumes normally after a few seconds

• Think Mode Only: Feature applies exclusively to Think mode (Claude Opus 4.6) for complex compliance work

Why This Helps:

• Work through comprehensive gap analyses, multi-framework mappings, and policy reviews without interruption

• No need to restart conversations or lose context during extended compliance discussions

• Reduces token costs for long-running AI interactions while maintaining conversation quality

How It Works:

When your Think mode conversation approaches Claude's context window limit, the system automatically invokes a summarization process in the background. You'll see an amber "Compacting our conversation..." indicator with a subtle progress bar. After a few seconds, the chat continues normally with the earlier context preserved as a summary. This allows indefinite conversation length while keeping the most relevant recent exchanges in full detail.

Free Tier Conversation Length Limit

Free plan users now have a 300,000 token per-conversation limit to ensure consistent performance across all users.

What's New:

• Per-Conversation Cap: Each individual conversation limited to 300k tokens (roughly equivalent to 200,000 words)

• Clear UI Card: When you reach the limit, a polished blue gradient card appears with a "Start New Conversation" button

• Blocked Input: Chat input automatically disables to prevent futile retries

• Premium Unlimited: Plus and Pro users enjoy 1 million token conversations (model context window limit)

Why This Helps:

• Prevents runaway token consumption in a single conversation

• Encourages starting fresh conversations for new topics

• Your previous messages are saved—just start a new chat to continue with a fresh context

Improved Context Limit Experience for Paid Users

Paid users who reach the AI model's 1 million token context window now see a clean, helpful interface instead of error messages.

What's Improved:

• Polished Blue Card: Clean gradient card replaces the old red error banner

• "Start New Conversation" Button: Prominent action that navigates you to a fresh chat

• Helpful Hint: Reminder that all previous messages are saved

• No Generic Upgrade Prompts: Business users (highest tier) no longer see misleading upgrade buttons when hitting usage limits

Detected Framework Badge

AI responses now show which compliance frameworks were detected and consulted, providing transparency into how the AI grounds its answers.

What You'll See:

• Collapsible Badge: Above each AI message, a "Detected framework >" label appears when frameworks are identified

• Expandable View: Click to see framework name badges (e.g., "ISO 27001", "SOC2", "GDPR")

• Localized Display: Framework names appear in your selected interface language

• Claude-Style Design: Clean, familiar disclosure pattern outside the message bubble

This helps you verify which knowledge bases the AI consulted for each response and understand the context behind compliance advice.

Improvements

Better Multi-Chapter Document Generation

Document generation now handles multi-turn requests (like generating policies chapter-by-chapter) more consistently and reliably.

What's Improved:

• Continuation Pattern Detection: Messages like "chapter 2", "next section", or "continue" are now recognized as document generation requests

• Format Keyword Awareness: Requests mentioning "docx", "pdf", "markdown", ".md", "file", or "document" never get misclassified as simple questions

• Longer Context Window: Generation context expanded from 5 to 10 messages to support extended multi-chapter conversations

• Enhanced Detection Prompt: AI better understands multi-turn document generation patterns and continuation cues

This fixes the issue where follow-up messages in multi-chapter policy generation (e.g., "now chapter 2") would sometimes be treated as questions instead of generation requests.

Faster Framework Detection

Framework detection is now significantly faster and more efficient, using fewer tokens while maintaining the same reliability.

What's Improved:

• Scoped Scanning: Detection now scans only user messages and the most recent AI response instead of entire conversation history

• Reduced Token Usage: Less token consumption means more room for your actual compliance queries within plan limits

• Same Accuracy: Detection quality remains unchanged—still catches ISO 27001, SOC2, GDPR, and all supported frameworks

• Faster Processing: Reduced scanning scope means quicker response times, especially in long conversations

This behind-the-scenes optimization makes the platform more efficient without changing how you interact with framework detection badges.

Smarter Framework Detection

Framework detection is now more efficient and accurate, using fewer tokens while maintaining the same reliability.

What's Improved:

• Token Optimization: Detection logic now scans only user messages and recent assistant responses instead of entire conversation history

• Reduced Costs: Less token consumption means more room for your actual compliance queries within plan limits

• Same Accuracy: Detection quality remains unchanged—still catches ISO 27001, SOC2, GDPR, and all supported frameworks

• Faster Processing: Reduced scanning scope means slightly faster response times

This behind-the-scenes optimization makes the platform more efficient without changing how you interact with framework detection badges.

Plan-Aware Limit Messaging

Usage limit messages now adapt to your subscription tier, eliminating confusing or misleading prompts.

What's Improved:

• No Upgrade Prompts for Business Users: Highest-tier users no longer see upgrade buttons when hitting session token limits

• Clearer Free-Tier Guidance: Free plan messages now clearly explain the 300k per-conversation limit and session-based quotas

• Blocked Input Placeholder: Changed from "upgrade or start new chat" to "please start a new chat to continue" (accurate for both free context limits and paid model limits)

• No Approaching-Limit Toast for All Users: Removed generic "Consider upgrading" from 80% usage warning toast

Better File Upload Experience for Free Users

Free tier file upload limits are now clearer and more helpful.

What's Improved:

• Distinct Unauthenticated State: Unsigned-in users see "Please sign in to upload files" instead of misleading limit-reached messages

• UTC Month Boundaries: Upload quotas now reset at consistent times regardless of your timezone

• Clear Remaining Count: Tooltip shows "X of 10 uploads remaining this month" for free users

• Upload Button Enabled When Signed Out: Button stays active so you can click and see the sign-in prompt

Scalable File Processing Timeouts

File processing now adapts timeout limits based on the number of files you upload, reducing false timeout errors.

What Changed:

• Dynamic Scaling: Base 3-minute timeout + 30 seconds per additional file (capped at 8 minutes)

• Better Multi-File Handling: Uploading 8 PDFs now gets 6.5 minutes instead of the old fixed 3 minutes

• Clearer Error UI: File processing errors now display as warm amber cards instead of aggressive red alerts

• Helpful Hints: "Uploading fewer files at a time can help" instead of unhelpful "try smaller files" messages

Advanced Data Protection Toggle Alerts

When you toggle Advanced Data Protection Mode on or off, you now see clear popups explaining the data handling implications.

What's New:

• Activation Info: Popup confirms AI switches to Mistral (EU-based) with zero retention and all data stays in EU

• Deactivation Warning: Amber alert explains that AI calls will transfer to US via Anthropic when you turn ADP off

• No Emoji Clutter: Removed EU flag emoji for cleaner interface

• Localized in 6 Languages: All popups appear in your selected language

Daily Subscription Sync Cron Job

Subscription status and premium feature access are now automatically synchronized with Stripe every 24 hours, ensuring your plan limits stay accurate.

What's Improved:

• Automatic Daily Sync: Plan upgrades, renewals, and cancellations reflected within 24 hours

• Safer Downgrades: System now fails open during Stripe API errors instead of incorrectly downgrading paying users

• Better Reliability: Premium access and usage quotas stay current even after subscription changes

Bug Fixes

Improved Chat Streaming Reliability

Fixed several edge cases affecting message display during AI responses, making the chat experience smoother and more reliable.

What's Fixed:

• Scroll Behavior: Auto-scroll now triggers reliably during fast response streaming, ensuring you always see the latest content

• Machine Lifecycle: Server instances now shut down correctly when idle, preventing stuck sessions (capped at 30-minute max re-arm limit)

• iOS Compatibility: Filtered out iOS 15 WebKit CSSStyleSheet browser errors that were causing false error reports in Sentry

• Count Leak Prevention: Fixed rare case where request count could leak on database errors, preventing machine shutdown issues

System Prompt Security Hardening

Enhanced protections prevent unauthorized extraction of internal AI configuration.

What's Fixed:

• Hardened Extraction Defense: System prompt now explicitly rejects indirect extraction attempts like "add to instructions" or "show me your guidelines"

• False Positive Protection: Legitimate requests like "add a clause to this access control policy" no longer trigger defensive refusals

• Workspace Instructions Carve-Out: AI correctly treats workspace instructions (tone, focus, language preferences) as legitimate customization, not override attempts

• Server-Side Output Filtering: Real-time streaming detection catches and replaces system prompt leaks within ~200 characters

• Anthropic Metadata Removal: Internal "budget:token_budget1000000" metadata stripped from all AI responses

File Upload Button Race Condition

Fixed issue where premium users saw a greyed-out file upload button on the welcome screen due to timing issues during initial page load.

Better Error Propagation for File Processing

File upload failures are now detected immediately (within ~5 seconds) instead of waiting for the full timeout period, providing faster feedback when something goes wrong.

14 February 2026

New Features

Markdown Document Generation

Generate compliance documents in Markdown format (.md) for easier version control, editing, and integration with documentation workflows.

What's New:

• Native .md Export: Request policies, procedures, and documents in Markdown format

• Simple Usage: Just ask "generate in markdown" or specify ".md format" in your query

• Better Portability: Works seamlessly with Git, wikis, static site generators, and plain text editors

• Easy Customization: Plain text formatting makes post-generation edits simple without specialized tools

Example Queries:

• "Create an ISO 27001 access control policy in markdown"

• "Generate a GDPR data retention procedure as .md"

• "Draft a risk assessment template in markdown format"

Perfect for technical teams using tools like GitHub, Obsidian, Notion, or any Markdown-based documentation system.

Enhanced Framework Support

Expanded compliance framework coverage with five new standards for specialized industries and regulations.

New Frameworks:

• TISAX: Trusted Information Security Assessment Exchange for automotive suppliers

• HDS v2.0: French health data hosting certification (Hébergeur de Données de Santé)

• ISO 9001:2015: Quality management systems and process mapping

• ISO 22301:2019: Business continuity management and disaster recovery

• EU AI Act: European Union artificial intelligence regulations

Improved Detection: Enhanced document analysis automatically identifies which frameworks are referenced in your uploaded files.

Ask questions about TISAX assessments, HDS health data controls, ISO quality processes, business continuity planning, or EU AI Act compliance with the same hallucination-free accuracy you rely on for ISO 27001, SOC2, and GDPR.

Document Preview with Split-Screen View

Preview generated documents directly in your chat without downloading. The new split-screen interface lets you review policies, procedures, and compliance documents instantly alongside your conversation.

What's New:

• In-Chat Preview: Click the preview button (eye icon) to open documents in a resizable split-screen panel

• Auto-Preview: Most recent document opens automatically for immediate review

• Copy as Markdown: One-click copy of full document content to clipboard in Markdown format

• Quick Download: Download button in preview header saves files to your device

• Format Support: Previews DOCX (rendered as HTML), Markdown, HTML, and plain text files

• Resizable Panels: Drag the divider to adjust chat and preview sizes

No more downloading files just to check content. Preview, review, copy, and download all from one interface.

Multilingual Interface (6 Languages)

Platform now supports six languages with full UI localization for European teams and multinational organizations.

Available Languages:

• English (default)

• German (Deutsch)

• Dutch (Nederlands)

• French (Français)

• Italian (Italiano)

• Spanish (Español)

Features:

• Full UI Translation: All menus, buttons, messages, and interface elements appear in your selected language

• Persistent Setting: Language preference saves to your profile across devices

• Easy Switching: Change anytime in Settings → Language

• No Impact on AI: AI conversations support any language regardless of UI setting

Perfect for multinational compliance teams and European consultancies. To change your language, go to Settings and select your preference.

Report Issue Button

New feedback mechanism makes it easier to report problems and share suggestions.

What's New:

• Dedicated Button: "Report Issue" option in chat feedback menu and user dropdown

• Auto-Context: Recent errors (within 30 minutes) automatically captured and pre-filled

• Pre-Filled Email: Your email address included for faster support response

• Streamlined Experience: Opens feedback form with one click—no extra dialogs

Improvements

Improved Document Naming

Generated files now use intelligent, readable filenames with better handling of custom names and special characters.

What's Improved:

• AI-Derived Names: Documents automatically named based on content when you don't specify a filename

• Custom Name Support: Specify your own filenames in queries (e.g., "Generate a policy named 'Access-Control-2026'")

• Smart Sanitization: Automatic cleanup of invalid characters while preserving readability (kebab-case)

• Duplicate Handling: System prevents filename conflicts automatically

• Accent Support: Correctly processes filenames with accented characters (é, ñ, ü, etc.)

No more generic "document_12345.pdf" names. Your policies and procedures now use meaningful, context-aware filenames.

Free Tier File Upload Limit

Free plan users can now upload up to 10 files per month to try multi-document analysis features.

What Changed:

• Monthly Quota: 10 total file uploads per month for free tier users

• Clear Display: Interface shows remaining upload quota

• Premium Unlimited: Plus and Pro users continue to enjoy unlimited file uploads

Free tier limit helps you explore document analysis capabilities. For regular use, consider upgrading to Plus or Pro for unlimited uploads.

Enhanced Framework Detection Badge

AI messages now display which compliance frameworks were detected and consulted for each response.

What You'll See:

• Framework Badge: Collapsible badge above AI messages shows detected frameworks (e.g., "ISO 27001, SOC2, GDPR")

• Expandable Details: Click to see why specific frameworks were identified

• Localized Labels: Framework names appear in your selected interface language

• Loading Indicator: Shows framework detection in progress during response generation

This transparency helps you verify which knowledge bases the AI consulted and understand the context behind each response.

Improved Limit Messaging

Clearer communication when approaching or reaching usage limits across all plan types.

What's Improved:

• Plan-Aware Messages: Prompts clearly explain your current plan tier and limits

• Visual Indicators: Progress warnings appear before you hit limits

• Session Resets: Clear explanation of when quotas refresh (midnight UTC)

• Context Limits: Better messaging for conversation length boundaries (300k free, 1M paid)

Context limits refer to total conversation history and uploaded documents the AI processes. Start a new chat or upgrade when you reach limits.

Advanced Data Protection Activation Popups

Clearer notifications when toggling Advanced Data Protection Mode help you understand data handling implications.

What's Improved:

• Toggle Warnings: Disabling ADP shows clear notice that data processes in US via Anthropic

• EU Confirmation: Enabling ADP confirms data stays in EU with Mistral AI and zero retention

• Localized Messages: Notices appear in your selected language

• Trade-off Clarity: Better explanation of privacy vs. performance implications

Learn more in Advanced Data Protection Mode.

File Processing Timeout Improvements

File uploads now have scalable timeout limits based on file size for more reliable processing.

What Changed:

• Dynamic Timeouts: Processing time scales from 3 minutes (small files) to 8 minutes (large files)

• Better Error Messages: Clearer feedback when uploads exceed time limits

• Reduced Failures: Larger files less likely to timeout during processing

Chat Stability Enhancements

Improved chat interface behavior with better scroll handling, component lifecycle management, and response indicators.

What's Improved:

• Smooth Scrolling: Auto-scroll to latest messages works more reliably

• Component Stability: Reduced remount issues that caused UI flicker

• Loading Indicators: Clearer feedback during message processing

Bug Fixes

Anthropic Provider Failover

Enhanced automatic failover when Anthropic Claude experiences outages. System now includes visual banner notifications during provider switches and automatic recovery when primary provider returns.

During Anthropic outages, requests automatically route to OpenAI backup. A persistent banner keeps you informed of provider status.

10 February 2026

New Features

Thinking Mode Enhancements

We've upgraded our AI thinking modes with the latest Claude models for faster, more accurate compliance analysis.

What's New:

• Fast Mode (Claude Sonnet 4): Updated to use Claude Sonnet 4 for improved speed and quality on everyday compliance queries

• Think Mode (Claude Opus 4.5/4.6): Access to the most advanced reasoning models for complex multi-framework analysis, gap assessments, and detailed policy reviews

• Premium Feature: Think mode remains available exclusively on Plus and Pro plans

• Persistent Setting: Your mode preference is saved automatically and persists across sessions

Detected Framework Visibility

The platform now shows you which compliance frameworks it has detected in your uploaded documents and conversations, giving you better visibility into how the AI is contextualizing your work.

What's New:

• Framework Badge: A collapsible badge appears above AI messages showing which frameworks were detected (e.g., "ISO 27001, SOC2, GDPR")

• Expandable Details: Click the badge to see full detection details and why specific frameworks were identified

• Localized Labels: Framework names appear in your selected interface language

• Context Awareness: Helps you understand which knowledge bases the AI consulted for each response

This transparency feature makes it clearer which frameworks are driving each response, helping you verify that the AI is using the right context for your compliance work.

Expanded Framework Support

Our AI knowledge base continues to grow with improved detection and coverage for additional compliance standards.

New Framework Coverage:

• TISAX: Trusted Information Security Assessment Exchange for automotive industry suppliers

• HDS v2.0: French health data hosting certification requirements

• ISO 9001:2015: Quality management system requirements and process mapping

• ISO 22301:2019: Business continuity management and disaster recovery planning

• EU AI Act: European Union artificial intelligence regulations and compliance requirements

Improved Detection: Enhanced regex-based framework detection in uploaded documents, making it faster to identify which standards your files reference.

Improvements

Improved Conversation Length Limits

We've improved how the platform communicates conversation length limits, making it clearer when you're approaching context boundaries.

What's Improved:

• Free Plan (300k tokens): Visual indicators show when you're approaching the 300,000 token conversation limit, with clear prompts to start a new chat or upgrade

• Paid Plans (1M tokens): Context limit cards appear when approaching the 1 million token window, helping you manage very long compliance discussions

• Input Protection: Chat input is automatically disabled when limits are reached, preventing wasted attempts

• Clearer Messaging: Updated UI clearly explains current usage and what to do next

Multiple File Upload Enhancements

Building on our multi-file upload feature, we've refined the limits and user experience for batch document analysis.

What's Improved:

• Free Tier Limit: Free plan users can upload up to 10 files per month total (not per batch)

• Batch Upload: Premium users continue to enjoy up to 10 files per batch with no monthly limits

• Clear Quota Display: The interface now shows your remaining file upload quota for the month

• Size Limits: Convertible files (PDF, DOCX, XLS) have a 5MB per-file limit to ensure reliable processing

Perfect for consultants analyzing client policy sets, comparing multiple compliance documents, or performing batch gap analysis. Learn more about file uploads.

Enhanced Session-Based Rate Limiting

We've improved how chat limits work across all subscription plans, making rate limit messages clearer and more consistent.

What's Improved:

• Session-Based Tracking: Rate limits now clearly show your remaining messages and when your quota resets (midnight UTC)

• Warning Messages: Proactive warnings appear when you're approaching your daily limit

• Consistent Experience: Same clear messaging across free and paid plans for any rate-limited features

• Better Visibility: Your current plan and usage status are always visible in the interface

Free plan users get 10 messages per day. Plus and Pro users enjoy unlimited messaging. Compare subscription plans.

Advanced Data Protection Mode Notices

We've improved how Advanced Data Protection Mode communicates its data handling to help you make informed privacy decisions.

What's Improved:

• Toggle Alerts: When you disable Advanced Data Protection, a clear warning explains that data will be processed in the US

• EU Confirmation: When enabled, a notice confirms your data stays in the EU and uses Mistral AI with zero retention

• Localized Messages: All notices appear in your selected interface language

• Clearer Trade-offs: Updated messaging helps you understand the privacy vs. performance implications

Advanced Data Protection Mode ensures 100% EU processing with zero AI provider retention. Learn more about Advanced Data Protection.

7 February 2026

New Features

Document Preview and Download

Generated documents now open in an interactive preview panel directly in your chat, eliminating the need to download files just to review them. When the AI generates a policy, procedure, or compliance document, you can preview it instantly in a split-screen view alongside your conversation.

What's New:

• In-Chat Preview: Click the preview button (eye icon) on any generated document to open it in a resizable split-screen panel

• Auto-Open: The most recently generated document opens automatically so you can review it immediately

• Copy Content: Copy the full document content to your clipboard as Markdown with one click

• Quick Download: Download button in the preview header for saving to your device

• Format Support: Previews DOCX files (rendered as HTML), Markdown, HTML, and plain text documents

• Resizable Panel: Drag the divider to adjust chat and preview panel sizes to your preference

How to Use:

1. Generate a document by asking the AI (e.g., "Create an ISO 27001 access control policy")

2. When the document appears in chat, click the Preview button or wait for auto-preview

3. Review the document in the split-screen panel

4. Use the Copy button to grab content as Markdown or Download to save the file

5. Close the preview or resize the panels as needed

Troubleshooting: If a document fails to preview, you'll see a clear error message with a download option. For file formats that don't support preview, the download button appears automatically. Learn more in Previewing generated documents.

Multilingual Interface Support

ISMS Copilot now supports six languages, making compliance workflows accessible to teams across Europe and beyond.

Available Languages:

• English (default)

• German (Deutsch)

• Dutch (Nederlands)

• French (Français)

• Italian (Italiano)

• Spanish (Español)

What's New:

• Full UI Translation: All interface elements, menus, buttons, and messages appear in your selected language

• Persistent Setting: Language preference is saved to your profile and persists across devices

• Easy Switching: Change language anytime in Settings without affecting your existing conversations or workspaces

• Framework Support: AI conversations remain multilingual-capable regardless of interface language

To change your language, navigate to Settings → Language and select your preferred option. The interface updates immediately.

Perfect for multinational compliance teams, European consultancies, and organizations working across language boundaries. AI responses continue to support any language in conversations—the UI translation makes navigation and settings clearer in your native language.

Improved Document Naming

Generated documents now respect your custom filenames with better sanitization and duplicate handling.

What's Improved:

• Custom Names: Specify document names when requesting generation (e.g., \"Generate a policy named 'Access-Control-2026'\")

• Smart Sanitization: Filenames automatically cleaned to remove invalid characters while preserving readability

• Duplicate Prevention: System detects and handles duplicate names gracefully to avoid file overwrites

• Accent Support: Correctly handles filenames with accented characters (é, ñ, ü, etc.)

No more generic \"document_12345.pdf\" filenames. Your generated policies and procedures now use meaningful names that match your compliance documentation structure.

Example Queries:

• \"Generate an ISO 27001 access control policy named 'AC-Policy-v2'\"

• \"Create a GDPR data retention procedure as 'Retention-Procédure-2026.docx'\"

• \"Draft a risk assessment named 'Q1-2026-Risk-Assessment.pdf'\"

Improvements

Clearer Free Plan Usage Messaging

Free plan users now see clearer, more actionable messages when approaching or hitting usage limits.

What's Improved:

• Session-Based Clarity: Messages now clearly explain that limits reset daily at midnight UTC

• Upgrade Guidance: Clearer prompts show upgrade benefits when you hit the 10 messages/day limit

• Better Context: Usage messages reference your current plan tier and available quota

Free plan users get 10 messages per day. If you regularly hit this limit, consider upgrading to Plus for unlimited messages. Visit Subscription Plans and Pricing to learn more.

Enhanced Loading Indicators

The AI loading experience now displays which framework the system is consulting, providing better visibility into processing steps.

What You'll See:

• \"Consulting ISO 27001 knowledge base...\" when processing ISO 27001 queries

• \"Consulting SOC2 framework...\" for SOC2 questions

• Framework-specific messages for GDPR, NIST, DORA, NIS2, and other supported standards

This makes it clearer which frameworks the AI is referencing to answer your question, helping you understand the sources behind each response.

3 February 2026

New Features

Markdown Document Generation

Generate compliance documents directly in Markdown format for easier editing and integration with your documentation workflows.

What's New:

• Native Markdown Export: Request policies, procedures, and compliance documents in .md format

• Simple Syntax: Just ask "generate in markdown" or "create a policy in .md format"

• Better Portability: Markdown files work seamlessly with version control, wikis, and static site generators

• Easy Editing: Plain text formatting makes it simple to customize outputs without specialized tools

Markdown is perfect for technical documentation, GitHub repositories, and teams using tools like Obsidian or Notion. Try asking: "Generate an ISO 27001 access control policy in markdown format."

Example Queries:

• "Create a GDPR data retention policy in markdown"

• "Generate an incident response procedure as .md"

• "Draft a SOC2 risk assessment template in markdown format"

Markdown generation works alongside existing export formats (PDF, DOCX, tables, checklists) giving you more flexibility in how you use ISMS Copilot outputs. Learn more about all available formats in Request Specific Output Formats.

Extended Framework Coverage

ISMS Copilot's AI knowledge base now includes three additional compliance frameworks, expanding your ability to get reliable answers across more standards.

New Frameworks:

• ISO 9001:2015: Quality management system requirements, clause-by-clause guidance, and process mapping support

• ISO 22301:2019: Business continuity management system requirements, including business impact analysis, risk assessment, and continuity planning

• HDS v2.0 (Hébergeur de Données de Santé): French health data hosting certification requirements, control mappings, and compliance guidance

• TISAX: Trusted Information Security Assessment Exchange for automotive industry suppliers, including assessment levels and maturity requirements

• EU AI Act: European Union artificial intelligence regulations, risk classification guidance, and compliance requirements for AI systems

Example Queries:

• "How does ISO 9001 clause 8.5.1 map to ISO 27001 operational controls?"

• "Create an ISO 22301 business impact analysis template for a financial services firm"

• "What are the HDS v2.0 requirements for health data encryption?"

• "Generate a TISAX assessment level 2 readiness checklist"

• "What are the EU AI Act transparency obligations for high-risk AI systems?"

Enhanced Subscription Plans

We've restructured subscription tiers to better match your compliance workflow needs. The new plan structure offers clearer pricing, more flexible billing cycles, and a new unlimited option for teams.

New Plan Tiers:

• Free: 10 messages/day, basic features, ideal for exploring the platform

• Plus ($24/month or $240/year): Unlimited messages, Think mode access, multi-file uploads, priority support

• Pro ($100/month or $1,000/year): Team workspaces, advanced collaboration, custom integrations

What's Improved:

• Clearer Upgrade Prompts: When you hit usage limits, you'll see a streamlined upgrade dialog explaining plan benefits and pricing

• Annual Billing Savings: Save 17-20% by choosing yearly billing

• Better Usage Visibility: Improved UI shows your current plan, usage limits, and upgrade options

Incident Resolution

Usage Limits Issue Resolved

Between Thursday, February 6 at 4:00 PM UTC and Sunday, February 9 at 9:00 PM UTC, we experienced issues with usage limits as part of a backend migration. During this period, some users encountered incorrect rate limiting or quota enforcement that prevented normal platform usage.

What Happened:

• Cause: A migration to improve subscription synchronization introduced temporary issues with usage limit enforcement

• Impact: Some users hit incorrect rate limits or saw inaccurate quota warnings

• Resolution: The migration has been completed and all usage limits are now functioning correctly

If you continue to experience any issues with abnormal usage limits or quotas, please contact support immediately.

Improvements

1M Context Window for Claude Models

Claude Sonnet 4.5 and Opus 4.6 now support a 1-million token context window, allowing you to process significantly larger compliance documents and have longer conversations without losing context.

What's Improved:

• Massive Context Capacity: Analyze documents up to 1 million tokens in a single conversation—equivalent to roughly 700,000 words or 1,500 pages

• Extended Conversations: Maintain context across longer compliance discussions, complex multi-framework mappings, and detailed gap analyses without hitting context limits

• Better Document Analysis: Upload and analyze larger policy sets, audit reports, and evidence packages in a single query

• Seamless Experience: The expanded context window works automatically—no configuration needed

Improved Subscription Management

Subscription status and access levels are now more accurate and reliable. A new daily background job automatically synchronizes your Stripe subscription with the platform, ensuring your plan limits and premium features are always up-to-date.

What's Improved:

• Automatic Sync: Daily reconciliation between Stripe and ISMS Copilot ensures your subscription status, message quotas, and access levels stay current

• Reliable Premium Access: Plan upgrades, renewals, and cancellations now reflected more quickly in the platform

• Better UI Accuracy: The top bar and settings now show correct plan information and usage limits

Stricter File Upload Limits

File upload size limits have been adjusted to prevent processing failures and improve reliability.

What Changed:

• New Limits: PDF, DOCX, and XLSX files now have a 5MB maximum (reduced from 10MB)

• Unchanged: TXT, CSV, and JSON files remain at 10MB

• Clear Feedback: The drag-and-drop interface now shows specific size limits for each file type

• Better Error Messages: You'll see specific error messages if your file exceeds the limit before uploading begins

These changes prevent stuck uploads and processing delays. Learn more in Uploading and Analyzing Files.

AI Provider Automatic Failover

ISMS Copilot now automatically switches to a backup AI provider when the default provider (Anthropic Claude) experiences outages or disruptions. The system monitors provider health in real-time using a circuit breaker pattern—tracking errors in a sliding window and automatically routing requests to OpenAI when Anthropic becomes unavailable.

What's New:

• Automatic Detection: Circuit breaker tracks 5xx errors, 529 rate limits, and network failures to detect provider issues

• Seamless Failover: Requests automatically route to OpenAI backup when primary provider is down

• Visual Notification: Persistent banner appears when failover is active, keeping you informed of provider status

• Automatic Recovery: System probes primary provider and switches back when it recovers

• No Disruption: Your work continues uninterrupted—failover happens transparently in the background

How It Works:

• The circuit breaker monitors failures during chat and document operations

• When errors exceed the threshold in the sliding window, the circuit "opens" and routes traffic to backup

• Periodic recovery probes test the primary provider to detect when it's healthy again

• Once primary recovers, the circuit "closes" and normal routing resumes

Provider Selection: Failover affects only users on the default provider path. If you explicitly select a specific model (OpenAI, Mistral, Gemini, Grok) in your settings, your selection is respected and bypasses automatic failover.

1 February 2026

New Features

Copy Button for AI Messages

Quickly copy AI responses with a single click. Every AI message now includes a dedicated copy button alongside feedback options, making it faster to extract compliance content for your policies, procedures, and reports.

What's New:

• One-Click Copying: Copy icon appears on every AI message for instant copying

• Integrated Feedback: Copy button positioned alongside thumbs-up/down feedback controls

• Streamlined Workflow: No more manual text selection—click to copy entire responses

Perfect for extracting control implementations, risk assessments, policy language, and compliance guidance directly from conversations. The copy function captures the full message content in a format ready to paste into your documentation.

Enhanced Issue Reporting

Report problems and provide feedback faster with streamlined reporting tools throughout the platform.

What's New:

• Pre-Filled Error Context: The "Report Issue" option in your user menu now automatically captures recent errors (within 30 minutes) and pre-fills your email for faster support

• Help & Feedback Button: New button in the left sidebar footer opens a feedback form with your email pre-filled

• Message-Level Feedback: Thumbs-up/down buttons on AI messages link directly to the feedback form with context

• Streamlined Experience: No dialogs or extra steps—hover and click to report

Whether you encounter a technical issue or want to suggest an improvement, the enhanced reporting tools make it easier to share feedback and get help.

File Upload Reliability Improvements

File uploads are now more reliable with stricter validation and clearer error messaging to prevent stuck uploads and processing delays.

What's Improved:

• Stricter Size Limits: PDF, DOCX, and XLSX files limited to 5MB; TXT, CSV, and JSON limited to 10MB to prevent processing failures

• Client-Side Validation: Files are checked before upload with clear error messages if they exceed limits

• Better Progress Feedback: Individual progress cards show upload status for each file in multi-file uploads

These improvements reduce upload failures and make it clearer when files cannot be processed due to size constraints. Learn more in Uploading and Analyzing Files.

January 2026

New Features

Multiple File Upload

Upload up to 10 files at once for faster multi-document analysis. Select multiple files from the file picker or drag-and-drop them directly into the chat input area.

What's New:

• Multi-Select: Click the paperclip icon and select up to 10 files at once, or drag multiple files anywhere in the chat input

• Individual Progress: Each file shows its own upload progress card with size and status

• Easy Management: Remove individual files by hovering over uploaded cards and clicking the × icon

• Visual Feedback: Drop overlay shows supported file types and limits when dragging files

Supported file types remain the same: PDF, DOC, DOCX, XLS, XLSX, TXT, CSV, JSON. Each file must be under 10MB.

How to Use:

1. Click the paperclip 📎 icon and select up to 10 files, or drag-and-drop multiple files into the chat area

2. Watch individual progress cards appear for each file

3. Remove any file by hovering and clicking the × icon

4. Once all uploads complete, send your message for AI analysis

Perfect for gap analysis across multiple policies, comparing documents side-by-side, or batch analyzing compliance evidence. Learn more in Uploading and Analyzing Files.

Fast and Think AI Modes

Choose between two AI processing modes to match your workflow: Fast mode for quick answers or Think mode for in-depth analysis.

What's New:

• Fast Mode: Get immediate responses for straightforward compliance questions

• Think Mode: Activate deeper reasoning for complex multi-framework mappings, risk assessments, or detailed policy reviews (available on Plus and Pro plans)

• Easy Switching: Toggle between modes directly in the chat interface based on your current task

Use Fast mode for day-to-day queries like control lookups or clause references. Switch to Think mode when you need comprehensive analysis, document reviews, or nuanced compliance guidance across multiple frameworks.

PII Reduction Mode

Automatically redact personally identifiable information before sending data to AI processing. When enabled, the system detects and removes names, email addresses, phone numbers, and company names.

What's New:

• Privacy Shield: A green shield icon appears in the chat when active

• Automatic Detection: Pattern-based redaction for common PII types

• Simple Toggle: Enable or disable in Settings → Privacy

Ideal for processing client documents with employee names, policies with contact information, or compliance evidence containing personal data. Works across all chat interfaces. Learn more in Data Privacy & GDPR Compliance.

Google Gemini AI Integration

ISMS Copilot now supports Google Gemini as an additional AI model option, giving you more choices for how compliance queries are processed.

What This Means:

• More Flexibility: Gemini joins OpenAI and Mistral as available AI providers

• Backend Integration: Model is integrated and ready for use (UI selection controls coming soon)

• Same Quality Standards: Gemini responses follow the same hallucination-free, framework-grounded approach

Model selection will be available in the Settings menu in an upcoming release. For now, ISMS Copilot automatically chooses the best model for your query.

Assistant Personality Improvements

The AI assistant now provides a warmer, more helpful experience with smarter handling of legal disclaimers and more proactive guidance.

What's Changed:

• Warmer Professional Tone: Responses now feel like working with a trusted advisor—clear and professional without being cold or robotic

• Natural Communication: Prefers prose over bullet points for more readable, conversational guidance

• More Proactive Editing: When reviewing documents, the assistant now provides direct fixes plus actionable checklists to help you improve compliance materials faster

• Context-Aware Legal Disclaimers: "Not legal advice" warnings now appear only when you're asking about legal topics like fines, contracts, or regulatory penalties—not during technical compliance work

• Clearer Uncertainty Handling: When the assistant isn't certain about something, it tells you directly rather than hedging

Data Retention Warning

A confirmation dialog now appears when reducing data retention settings, helping prevent accidental deletion of workspace data.

What Changed:

• Protection Against Accidents: Confirmation required when switching from "Keep Forever" to shorter retention periods

• Clear Impact: Dialog explains what data will be affected before you confirm

• GDPR Alignment: Helps manage data lifecycle requirements while preventing unintended data loss

Improvements

Quick Access to Product Updates

You can now access the Product Changelog directly from your user menu (click your avatar → Help Center → Product Changelog). Stay informed about the latest features without leaving the app.

Enhanced Outage Communication

Improved status banners now provide clearer information during AI provider disruptions, including partial outages and recovery status.

Better Error Messages

Chat errors now display clearer explanations for rate limits, timeouts, and quota issues, making it easier to understand what happened and how to proceed.

December 2025

New Features

PII Reduction Mode

ISMS Copilot now offers automatic personally identifiable information (PII) redaction to enhance privacy when processing compliance documents and queries. When enabled, the system automatically detects and redacts names, company names, email addresses, and phone numbers before sending data to AI processing.

What's New:

• Privacy Shield Indicator: A green shield icon appears in the chat input when PII reduction is active, providing clear visual confirmation

• Automatic Pattern Detection: Redacts common PII patterns including personal names, email addresses, phone numbers, and organization names

• User Control: Simple toggle in Settings to enable or disable PII reduction

• Confirmation Dialog: When enabling PII reduction, a popup explains the feature's limitations and pattern-based approach

How to Enable:

1. Navigate to Settings → Privacy or Data Protection

2. Toggle "Enable PII Reduction"

3. Review and confirm the limitations popup

4. Look for the green shield icon in your chat input to confirm activation

Use Cases:

• Processing client documents containing employee or customer names

• Analyzing policies with contact information

• Working with compliance evidence containing personal data

• Additional privacy layer when using Advanced Data Protection Mode

PII reduction works across all entry points including the home chat and workspace conversations. Learn more in Data Privacy & GDPR Compliance.

OpenAI GPT-5.2 Backend Integration

ISMS Copilot now uses OpenAI's GPT-5.2 model as the default AI provider on the backend, offering enhanced reasoning capabilities and compliance analysis quality.

What's New:

• Default Provider: GPT-5.2 is now the default AI model (Update: As of early 2026, Anthropic Claude is the default provider for chat replies, with GPT used for detection and Grok for document generation.) powering all conversations

• Backend Integration: Claude Opus and other AI providers are integrated on the backend, with model selection UI coming in 2026

• Improved Accuracy: Enhanced framework interpretation and control mapping precision

• Same Experience: You'll continue using ISMS Copilot as before, with better AI performance under the hood

Dynamic Framework Knowledge Injection

ISMS Copilot now automatically injects relevant framework clauses and control requirements directly into responses based on your query context.

How It Works:

• Automatic Detection: System identifies which frameworks (ISO 27001, SOC2, NIST, GDPR, etc.) are relevant to your question

• Clause Integration: Embeds specific control requirements, clauses, and implementation guidance in responses

• Context-Aware: Adapts based on workspace settings and conversation history

• Multi-Framework Support: Covers ISO 27001/27701/42001, SOC2, NIST CSF, GDPR, DORA, NIS2, and more

Example Use Cases:

• Ask "How do I implement access control?" → Response includes ISO 27001 A.9 controls, SOC2 CC6 criteria, NIST PR.AC references

• Query about data retention → Auto-includes GDPR Article 5(1)(e) and relevant ISO 27001 controls

• Risk assessment questions → Embedded with ISO 27001 Clause 6.1.2 requirements

Premium Upgrade Dialog for Rate Limits

Users on the free plan now see a clear, user-friendly upgrade prompt when reaching the daily message limit (10 messages/day). The new full-screen dialog explains the benefits of upgrading to Plus ($24/month) for unlimited access and can be dismissed to explore subscription options.

Improvements

Enhanced Agentic Loading Experience

The loading experience now provides richer visual feedback about what ISMS Copilot is doing behind the scenes. Instead of a simple spinner, you'll see an animated 3×3 dot grid with contextual messages that rotate as the system processes your request.

What You'll See:

• Animated dot grid visualization during processing

• Rotating status messages like "Analyzing your question...", "Consulting framework knowledge...", "Preparing response..."

• Clear indication when the system is actively working on complex queries

This provides even clearer understanding of the AI's multi-step processing workflow.

Cleaner AI Responses

AI-generated responses no longer include [Part X/Y] markers or segmentation indicators. Long responses now appear as continuous, clean text without technical artifacts.

Benefits:

• More professional output suitable for direct use in compliance documentation

• Easier to copy-paste responses into policies and procedures

• Cleaner reading experience for complex, multi-part answers

Relaxed System Prompt Guardrails

Workspace-level system prompts now accept legitimate context and custom instructions while maintaining ethical boundaries. The updated guardrails:

• Accept legitimate workspace instructions, client context, and intellectual property

• Continue to reject unethical requests or harmful content

• Provide more flexibility for consultants managing multiple client workspaces

You can now provide more detailed workspace context without triggering false-positive content filters. Learn more about workspace customization in Organizing Work with Workspaces.

Expanded Framework Coverage

Significantly enhanced coverage and accuracy for multiple frameworks with updated knowledge bases and reference tables:

• ISO 27001:2022: Updated control mappings, implementation guidance, and audit evidence requirements

• ISO 27701:2025: Enhanced privacy control guidance with latest standard updates

• ISO 42001:2023: Comprehensive AI governance framework knowledge including all controls and requirements

• SOC2: Refreshed trust services criteria mappings

• NIST CSF: Updated with latest framework versions and subcategories

These updates ensure you're working with the most current framework interpretations and can generate audit-ready documentation.

Extended Chat Timeout

Chat polling timeout has been extended to 10 minutes, significantly reducing instances of stuck or incomplete responses. This improvement provides more reliable chat interactions, especially for complex compliance queries.

Enhanced Help Center Integration

All in-app help links now direct to help.ismscopilot.com for consistent access to documentation, tutorials, and support resources.

Improved Footer Messaging

The footer now includes a link to the user guide with the message "ISMS Copilot is good but you remain the captain" — reinforcing that the tool accelerates your compliance work while you maintain oversight.

Bug Fixes

Consistent PII State Management

Fixed an issue where PII reduction settings could become inconsistent across different parts of the application. The toggle state now reliably synchronizes across all chat interfaces.

Response Formatting

Resolved backend issues that caused [Part X/Y] markers to appear in responses, ensuring clean output for all queries.

Chat Loading Spinner

Fixed infinite loading spinner issues that could occur during chat interactions. The interface now correctly indicates when responses are processing or complete.

Message Concatenation Errors

Resolved backend concatenation errors that were previously visible to users. Chat responses now display cleanly without technical error messages.

Workspace Message Display

Fixed messaging display issues in workspace-specific contexts, ensuring consistent chat behavior across all workspaces.

November 2025

New Features

Advanced Data Protection Mode

You now have complete control over where your AI conversations are processed and how long AI providers retain your data.

Two Processing Options:

• Default Mode (OFF): AI processing in the United States via xAI/OpenAI with 30-day retention

• Advanced Data Protection (ON): 100% EU processing via Mistral AI with zero data retention

Key Benefits:

• User control over AI processing location (US vs EU)

• Zero retention option for maximum privacy

• Meets strict EU data sovereignty requirements

• Available to all users at no additional cost

• Simple toggle in Settings → Data Protection

Who Should Enable This:

• Organizations with mandatory EU data residency requirements

• Compliance consultants working with European clients

• Users handling highly sensitive compliance data

• Anyone who wants maximum privacy and zero AI provider retention

Related Documentation Updates:

• Updated Data Privacy & GDPR Compliance to reflect user control over AI processing location

• Updated Register of Processing Activities to document dual-mode AI processing

• Updated Security & Data Protection Overview with AI provider options

October 2025

New Features

Help Center & Support Resources

We've enhanced access to support resources directly within ISMS Copilot:

• Help Center submenu: Access frequently asked questions, tutorials, and support documentation from the user dropdown menu

• Improved navigation: Quickly find the help you need without leaving your workspace

Mistral AI Model Integration

You can now choose Mistral as an alternative AI model for your compliance queries. This gives you more flexibility in how you interact with the platform and allows you to select the model that best fits your specific use case.

Refer and Earn Program

Introduce ISMS Copilot to other compliance professionals and earn rewards. The new "Refer and Earn" link in your user menu makes it easy to share the platform and track your referrals.

Improvements

Enhanced AI Response Quality

We've refined the AI's system prompt to deliver more concise and natural responses while maintaining accuracy for compliance-specific queries.

Payment Processing Updates

Updated Stripe integration for more reliable payment processing and improved billing experience.

Bug Fixes

Chat Typing Animation

Fixed an issue where the typing animation in chat felt sluggish. Messages now appear more smoothly and responsively.

Document Generation Reliability

Resolved several bugs affecting the document generation process, improving overall reliability when creating policies, procedures, and compliance documentation.

How to Stay Updated

We continuously ship improvements to ISMS Copilot based on user feedback and compliance framework updates. To make the most of new features:

• Bookmark this page and check back regularly for updates

• Watch for in-app notifications about new features

• Join our community to discuss new releases and share feedback

Getting Help

If you have questions about any of these updates or need assistance:

• Visit the Help Center from your user menu for tutorials and FAQs

• Contact support for technical assistance

• Check related articles for guides on specific features